David Schultz 9fd7a48db0 Cut out the gordian handling of subnormals by bit fiddling, and
instead use the FPU to convert subnormals to normals.  (NB: Further
simplification is possible, such as using the FPU for the rounding
step.)

This fixes a bug reported by stefanf where long double subnormals in
the Intel 80-bit format would be output with one fewer digit than
necessary when the default precision was used.
2005-01-18 18:44:07 +00:00
..
2004-12-21 10:49:29 +00:00
2005-01-10 10:54:06 +00:00
2004-09-09 10:54:08 +00:00
2004-11-04 08:25:03 +00:00
2004-12-21 09:08:06 +00:00
2003-06-14 19:32:52 +00:00
2004-12-21 09:33:47 +00:00
2004-12-21 10:16:04 +00:00
2004-12-21 10:16:04 +00:00
Use
2003-06-06 13:46:55 +00:00
2004-07-01 18:20:57 +00:00
2004-12-21 10:49:29 +00:00
2004-12-29 02:18:24 +00:00
2004-07-05 06:59:17 +00:00
2004-12-28 04:35:01 +00:00
2004-07-03 22:30:10 +00:00
Use
2003-06-06 13:46:55 +00:00
2004-12-21 10:49:29 +00:00
2004-12-21 12:05:11 +00:00
2004-12-21 10:16:04 +00:00
Use
2003-06-06 13:46:55 +00:00
2004-12-21 10:49:29 +00:00
2004-12-21 10:16:04 +00:00
2003-08-18 15:25:39 +00:00
2004-12-21 10:49:29 +00:00
2004-12-21 10:16:04 +00:00
2004-07-05 06:53:34 +00:00
2005-01-11 20:50:51 +00:00
2004-12-21 10:49:29 +00:00
2004-12-21 09:33:47 +00:00
2004-12-21 12:05:11 +00:00
2004-12-21 10:16:04 +00:00